Project 4 / Scene Recognition with Bag of Words
The project is implemented With two different feature selectors: tiny image, and bag of SIFT; and two classifier: nearest neighbor, and SVM.
The following experiments are ran on 1500 images with num_train_per_cat at 100.
Results in a table
 |
 | |
| tiny image + SVM | tiny image + NN | |
 |
 | |
| SIFT + NN | SIFT + SVM |